Découpage : page précédée de U+2060, titres coupés sur plusieurs lignes, saut de page dans la table des matières (trois tests). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
44 lines
1.7 KiB
Python
44 lines
1.7 KiB
Python
"""Le corpus versionné de la base de connaissances : présent, propre, sans rien d'interne (spec §3, §6)."""
|
|
import re
|
|
import unittest
|
|
from pathlib import Path
|
|
|
|
from base_connaissances.decoupage import lire_corpus
|
|
|
|
CORPUS = Path(__file__).resolve().parents[1] / "base_connaissances" / "corpus"
|
|
STATIONS = ["BREITHORN", "CLARIDEN", "FLORIDA", "JUNGFRAU", "LENZSPITZE", "MICHIGAN", "NADELHORN", "NEBRASKA",
|
|
"NIEDERHORN", "NORDEND", "SNCF01", "SUSTENHORN", "WILDSTRUBEL", "VERMONT"]
|
|
INTERDITS = ["python3", "agent-tools", ".py", "Marc", "Barcelone", "Vaud", "GN-1", "GN-5", "seuil provisoire",
|
|
"seuils provisoires"]
|
|
|
|
|
|
class TestCorpus(unittest.TestCase):
|
|
def test_corpus_present(self):
|
|
self.assertGreaterEqual(len(lire_corpus(CORPUS)), 60)
|
|
|
|
def test_titres_uniques_et_contenus_non_vides(self):
|
|
x = lire_corpus(CORPUS)
|
|
self.assertEqual(len({e.titre for e in x}), len(x))
|
|
self.assertTrue(all(e.contenu.strip() and e.source.strip() for e in x))
|
|
|
|
def test_taille_raisonnable(self):
|
|
trop_longs = [e.cle for e in lire_corpus(CORPUS) if len(e.contenu) > 8000]
|
|
self.assertEqual(trop_longs, [])
|
|
|
|
def test_rien_d_interne(self):
|
|
for e in lire_corpus(CORPUS):
|
|
texte = e.titre + "\n" + e.contenu
|
|
for s in STATIONS:
|
|
self.assertIsNone(re.search(rf"\b{s}\b", texte, re.I), f"{s} dans {e.cle}")
|
|
for i in INTERDITS:
|
|
self.assertNotIn(i.lower(), texte.lower(), f"{i} dans {e.cle}")
|
|
|
|
def test_instructions_presentes(self):
|
|
texte = (CORPUS.parent / "instructions.md").read_text()
|
|
self.assertIn("langue du client", texte)
|
|
self.assertLessEqual(len(texte), 4000)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|